How Real Estate Comps Are Chosen (And Why Bad Ones Ruin Valuations)

Real estate comps are chosen by finding recently closed sales that match a subject property on location, property type, size, age, and condition, then adjusting each one for its remaining differences. The selection step matters more than the adjustment step, because a bad comp set produces a wrong answer no matter how carefully you adjust it.

What is a real estate comp?

A comparable sale, or comp, is a closed transaction used as evidence of what a subject property is worth. Appraisers, agents, lenders, and automated valuation models all work the same way: find sales that a buyer would have considered a genuine alternative to your property, then reason from what they actually paid. The word doing the work is *alternative*. A comp is only valid if a buyer shopping for your property would have plausibly bought that one instead.

How are comps chosen?

Selection runs through a series of filters, tightest first, loosening only as needed to reach a usable count.

1. Geography

Location is the strongest filter and the least negotiable. The preferred order is:

1. Same subdivision, project, or block face 2. Same neighborhood with the same school attendance boundaries 3. Same market area within roughly 0.5 to 1 mile in suburban settings 4. Beyond that, only with an explicit reason and a location adjustment

Crossing a boundary is the most common way comp sets go wrong. A major road, a school district line, a city limit, a flood zone edge, or the border of a golf course community can create a 15 to 30 percent value break across two streets that look identical on a map.

2. Time

Recent sales reflect current conditions. A typical window is 3 to 6 months, extended to 12 months only in thin markets where nothing else has sold. Older comps need a market conditions adjustment, and if the adjustment is large, the comp is weak evidence.

3. Property type and ownership form

Match detached to detached, attached to attached, and match the ownership form. A fee simple townhouse in the same project as a detached home can be a reasonable comp because the buyer pool and cost structure overlap. A condominium with shared elements and a monthly assessment is a different product and usually does not comp to either.

4. Size

Living area is normally held within about 20 percent of the subject. The reason is that price per square foot is not linear: smaller homes command a higher price per foot and larger homes a lower one, so comping a 1,200 square foot house to a 2,200 square foot house injects error that no adjustment fully removes.

5. Age, style, and layout

Similar year built, similar story count, similar bed and bath configuration. A 4 bed and a 3 bed of identical square footage sell to different buyers.

6. Condition

This is the filter most commonly skipped, and it is the one that decides whether you produced an as-is value or an ARV. Comping a renovated home against dated sales understates it. Comping a gut job against renovated sales overstates it, often badly.

7. Sale type

Arm's length sales only. Foreclosures, auction sales, estate sales, and transfers between family members reflect conditions other than open market value and should be excluded or heavily discounted.

How do you adjust comps?

Once selected, each comp gets adjusted toward the subject. The direction is always the same and always confuses people at first:

• •Comp is superior to the subject, adjust the comp's price down. - Comp is inferior to the subject, adjust the comp's price up. You are answering: what would this comp have sold for if it had been my property? Adjustments come from market evidence where possible, such as paired sales of homes that differ only in one feature. In practice most investors use market-supported ranges for living area, garage, bath count, lot size, pool, and view, then adjust condition last and largest.

Worked example (illustrative numbers)

Subject: 3 bed, 2 bath, 1,600 square foot detached home, renovated, 2 car garage, no pool.

Comps A and B are the strongest evidence. Comp C required a large location adjustment, which is a signal that it should carry the least weight or be replaced. The supported value sits near $355,000. Watch what happens if you drop Comp A because it is "too small a garage" and keep only B and C without adjusting for location. You would read the raw prices, $368,000 and $381,000, and conclude the property is worth about $374,000. That is roughly $19,000 of error created purely by comp selection, before any adjustment math.

Why do bad comps ruin valuations?

Because valuation error compounds downstream. In investor underwriting, a wrong ARV moves your maximum offer, your loan sizing, your projected profit, and your refinance expectation all at once. The 70 percent rule multiplies ARV by 0.70, so an ARV that is $19,000 high raises your offer ceiling by about $13,300 while simultaneously removing $19,000 from your exit. Four failure patterns account for most bad comp sets:

Crossing a submarket boundary. The comp is close on a map but on the other side of a value break. This produces confidently wrong numbers because everything looks reasonable. Ignoring condition. The comp closed at a high price because it was fully renovated. If your subject is dated and you did not adjust for it, you have inherited someone else's finished-condition value. Chasing count over quality. Six mediocre comps are worse than three strong ones. Adding weak comps to reach a target count widens the range and lowers your confidence in every number in the set. Accepting a black box. If you cannot see which sales produced a value, you cannot audit it, and you certainly cannot defend it to a lender, a partner, or a seller.

How to audit a comp set in five minutes

1. Map them. Put every comp on a map with the subject. Look for roads, district lines, and water. Anything on the far side of a boundary needs a reason to stay. 2. Check the adjusted spread. If adjusted values are tight, the set is coherent. If they are still 15 percent apart after adjustment, the selection is wrong. 3. Look at the extremes. Open the highest and lowest comp and ask why each is where it is. Usually one is a renovation level you are not matching. 4. 5. Remove your best comp and re-run. If the value collapses, your conclusion rests on one sale.

Comps you can actually see

Resideline shows the comps behind every valuation and lets you adjust or remove them, rather than returning a number with no supporting evidence. Condition is estimated from listing photos, which is an estimate and not an inspection, and that estimate is what separates the as-is framing from the ARV framing. Estimates are frozen when a property lists and later graded against the real closing price, with results published on the public accuracy dashboard. To put comps to work, run a value with the free ARV calculator, pull the supporting sales into a CMA report you can hand to a seller or lender, and take the deal through the deal analyzer. For screening many addresses at once, ARV Runner handles bulk pulls. Live valuations currently cover 31 US states.